commit c407e30e texdraw: negative of unsigned int requires a cast

Probably not a big deal, but...
     I believe this commit to be incorrect.

It is not true that passing the negative of an unsigned int requires a cast.
Or if it is true, then something is wrong with the compiler.

If the function has a prototype (now required!) then arguments are promoted to
match the prototype.

fprintf() is a weird special case because it is a variadic function but nevertheless 
anything given a format specifier %d will be promoted to integer.

This is covered in section 6.5.2.2 of the ISO C standard (checked in draft dated 2007)

If for some reason there is a real problem here I would prefer to understand
what it is.   Did you get a compiler warning?  What compiler?
Do you have a test case where this commit actually makes a difference to the output?

If a change is necessary for some reason, I would rather consider
all of these structure fields to int rather than unsigned.

*diff --git a/src/term_api.h b/src/term_api.h* 
*index c836e19cd..f8898c0b5 100644* 
*--- a/src/term_api.h* 
*+++ b/src/term_api.h* 
@@ -290,7 +290,7 @@ typedef enum t_imagecolor { IC_PALETTE, IC_RGB, IC_RGBA } 
-    unsigned int xmax,ymax,v_char,h_char,v_tic,h_tic; 
+    int xmax,ymax,v_char,h_char,v_tic,h_tic; 


  best,
	Ethan

%% commit    c407e30e
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
@@ -410,11 +410,11 @@ TEXDRAW_graphics()
 	/* enforce bounding box */
 	fprintf(gpoutfile, "\\move (0 0) \\rmove (%d %d)\n",
 		    term->xmax, term->ymax);
     } else {
 	fprintf(gpoutfile, "\\move (0 0) \\rlvec (%d 0) \\rlvec (0 %d) \\rlvec (%d 0) \\ifill f:%0.2f\n",
-		    term->xmax, term->ymax, -term->xmax,
+		    term->xmax, term->ymax, - (int) term->xmax,
 		    TEXDRAW_background);
     }
 
     TEXDRAW_last_type = 0;
     TEXDRAW_type = 0;
@@ -814,11 +814,11 @@ TEXDRAW_fillbox(int style,
 
     // outline box using relative moves
     fprintf(gpoutfile, "\\move (%d %d)", x1, y1);
     fprintf(gpoutfile, "\\rlvec (%d %d)", width, 0);
     fprintf(gpoutfile, "\\rlvec (%d %d)", 0, height);
-    fprintf(gpoutfile, "\\rlvec (%d %d)", -width, 0);
+    fprintf(gpoutfile, "\\rlvec (%d %d)", - (int) width, 0);
     // the polygon is closed automatically by fill
     fprintf(gpoutfile, "\\ifill f:%0.2f\n", gray);
 }
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

_______________________________________________
gnuplot-beta mailing list
[email protected]
Membership management via: https://lists.sourceforge.net/lists/listinfo/gnuplot-beta